A Machine Learning Approach to the Classification of Engine Health

Engine quality is the fulcrum of effective transportation and determines the health of the car. However, there are several parameters that need to be taken into consideration while estimating engine health which makes it cumbersome and tedious to always do it manually using techniques including but not limited to compression tests, leak down tests etc. This is where machine learning can step in. Machine learning uses historical data for prediction or classification tasks and can be extremely useful for this task of classifying the engine quality as normal, or, one with faults. This can be an extremely useful tool to aid engine quality assessment either during testing phases or can be incorporated into the car for onboard diagnostics and early-stage detection and repairs making life comfortable for the user and increases the reputation of the automobile brand. In this paper, we therefore apply several machine learning techniques for classification of engine quality namely Logistic Regression, Artificial Neural Networks, Support Vector Machines, Decision Trees and Random Forest using several input features. We evaluate model performance using metrics such as training time, accuracy, precision, recall and model stability. We recommend the best model for this task of engine health classification and show that this is the way forward for intelligent maintenance services offered to the car owners in this digital age.